A Simple Distributed Relay-Assignment Algorithm
for Multi-Antenna Relay Networks
Chao Zhang, Lei Wang, Weidong Wang and Guo Wei
Dept. of Electronic Engineering & Information Science,
and Wireless Information Network Lab.
University of Science and Technology of China, Hefei, China.
Email: lockhead@mail.ustc.edu.cn
Abstract—In multi-antenna relay networks, how to assign source-
relay pairs is still a challenge. In this paper, we propose a simple
distributed relay-assignment algorithm for multi-antenna relay
networks, which exploits the ability of AOA (Angle of Arrival)
estimating in a multi-antenna terminal. The proposed algorithm
based on an opportunistic relaying model uses RTS packet to
estimate AOA of the source and CTS packet to estimate AOA of
the destination in potential relays, and then calculates the
interval angle of each node. The best relay is chosen to be with
the maximum interval angle between source-relay link and relay-
destination link. An outage analysis for the proposed algorithm is
provided within a limited area where available relaying nodes are
uniformly distributed. Finally, simulation results show that the
proposed algorithm offers great improvement on the reliability of
the relaying links and is more suitable for multi-antenna relay
networks than existing relay-assignment algorithms. The outage
analysis is also verified.
Keywords-cooperative relaying; relay-assignment
I. INTRODUCTION
Cooperative diversity is a set of techniques that exploit the
potential of spatial dispersed terminal/node antennas to
improve communication reliability. The basic idea is that relays
are assigned to help a source in forwarding its information to
its destination to achieve the spatial diversity. Various
cooperative diversities where each user or node is equipped
with single antenna were proposed and analyzed in [1]-[3] to
mimic the performance advantages of multi-antenna systems.
More recently, applying MIMO techniques to relay networks
has also come under consideration, i.e. [4], [5]. For multi-
antenna or MIMO relay channels where nodes in the network
can be deployed with multiple antennas, most previous studies
are aimed at obtaining information-theoretic limits and
diversity gains, just assuming the source-relay pair is prefixed.
In this work, we examine the problem of relay-assignment in
a multi-antenna relay network of multiple relays. As for single
antenna relay networks, selective relaying is proved to have
better outage behavior and higher average throughout than
conventional scheme in [6]. So how to select a best node to
relay effectively has gained much interest in research
community. Moreover, centralized selective relaying schemes
bring much system overhead into the cooperative networks and
the overhead will counteract partial relaying gains [7]. Recently,
few distributed relay-assignment algorithms are proposed for
single antenna relay networks. In [7], opportunistic relaying is
proposed to choose the relay with best instantaneous channel
condition, based on an 802.11b-like MAC protocol. For multi-
antenna relay networks, opportunistic relay algorithm will lead
to more consumed energy to obtain multiple CSIs (Channel
State Information) and complicated relay selecting criteria to
evaluate the whole state of a multi-antenna terminal. [8]
presented a Nearest-Neighbor algorithm to select the nearest
user to the source as the relay to extend the transmission
coverage. In [9], a fixed priority selection is proposed to
achieve full diversity. These two distributed relay-assignment
algorithms are not the optimal scheme in terms of relaying
performance or bandwidth efficiency. Although, all
aforementioned algorithms can be used for multi-antenna relay
networks, these do not exploit the potential of the multi-
antenna relay well.
Taking into account of the ability of AOA estimating in
multi-antenna terminals [10], we propose a simple distributed
relay-assignment algorithm based on a simplified opportunistic
relay scheme. The rationale behind the proposed algorithm is
that a node which is closest to the source-destination line
should be chosen as the relay to decrease path loss. Although,
this choice might not be the optimal in all scenarios, it is very
simple to implement in a distributed manner and can achieve
good performance as we will demonstrate later. The
source/destination broadcasts a RTS/CTS packet to all the
available multi-antenna nodes at the beginning of a data packet
transmission to estimate AOA of the source/destination.
According to the two estimated AOAs of both the source and
destination, each node calculates its interval angle between the
link of source to itself and the link of destination to itself. The
node with maximum interval angle broadcasts its flag packet to
inform other nodes it is the relay. After the relay is selected, it
can be used to forward information to the destination. Once the
relay is assigned, any cooperation scheme for multi-antenna
relay networks can be employed. We analyze the outage
probability for the proposed algorithm in Decode-and-Forward
scheme, and the expression for average outage probability is
derived. Simulations show that the performance of proposed
algorithm is better than that of methods of [8], [9] and Policy 1
in [7], and approaches the performance of Policy 2 in [7],
which is CSI based. However, for a relay with L antennas, CSI
based method needs L CSI-estimators, and the cost drastically
increases as the value of L increases. Our algorithm just needs
only one AOA-estimator all the time. Therefore, the proposed
This work was supported in part by National Basic Research Program of
China (973Program) , No. 2007CB310602 and by Intel Research Council ,
o.4507334166
Authorized licensed use limited to: BEIJING UNIVERSITY OF POST AND TELECOM. Downloaded on December 21, 2008 at 20:55 from IEEE Xplore. Restrictions apply.